Reproducibility: Sharing Your Research

Documentation

Just because someone has a copy of your data doesn't mean they'll be able to interpret it correctly. Metadata allows researchers to specify everything from what the labels in your table columns mean, to what were the precise settings you used when capturing an image. Make sure you are keeping track of the context necessary for others to use your data.

Dryad README Guidance
General guidance on what to include in a README file
rOpenSci's Best Practices for Project Metadata
A guide for those working in R
RDA Metadata Standards Directory
Different disciplines use different metadata standards for documenting their data; if your discipline has coalesced around one standard it's important to follow their lead.

Repositories

Reproducible research requires that you not only share your process, but also your underlying data. It's important to find a trustworthy open-access repository to host this data, because you don't want researchers to have to contact you every time they're interested in working with it.

UW-Supported Options

Dryad
Dryad is a generalist data repository offering a secure location for research data dissemination. UW is now a Dryad member and UW researchers can deposit their data in Dryad at no additional cost to them -- deposit costs are covered by the Libraries' membership fee. This is especially important for researchers with data sets that do not fit in any of the federal or subject repositories.
UW ResearchWorks
UW's open access institutional repository, which offers free archiving. ResearchWorks can mint Digital Object Identifiers (DOIs) for your work.
UW Data Storage Options
A list of UW storage options, a number of which provide open access

Disciplinary Options

re3data
Specific research fields tend to have their own repositories, and this website is a good starting point for finding one that will suit your needs

General Options

Zenodo
An EU-backed repository for data of all types
figshare
An online repository for data and visualizations that may or may not relate to publications

Licensing

Software and some types of data are subject to U.S. copyright law, so it's important to communicate to potential users the set of circumstances under which they may legally use the materials.

Software

Generally, UW faculty, staff, and students retain all rights in copyrightable materials they create. However, this is subject to a number of exceptions, one of which is for 'software of commercial value.' For this kind of software, copyright is assigned to the University. UW grants to the authors the right to distribute software of commercial value at no cost for any purpose the author chooses, provided such software is not already under management by CoMotion and provided such software is made broadly available. CoMotion should be consulted prior to distribution if there are commercialization goals for the software.

UW recommends that software be made openly available through a BSD license, and explicitly prohibits distribution through Apache or GPLv3 licenses.

Data

Although facts are not covered by U.S. copyright law, the organization of facts may be. Because of this, and because international law treats data differently in different jurisdictions, it's important to attach a license to your data as well as to your code.

UW recommends that you use a CC0 license for your data, but you have a number of options.

Summary of Options

	Software	Data	Creative Works
Open Source Licenses	✓
BSD	✓ (UW)
Apache	X (UW)
GPLv3	X (UW)
CC0	✓	✓	✓
Creative Commons Attribution		✓ (?)	✓
Open Data Commons		✓

Pre-registration

In some fields, such as psychology, it is becoming common to pre-register a study. This is a public statement committing to your plan in advance, before you gather data. This helps to prevent situations where the same data are used to generate and test a hypothesis.